NDUFAB1 Confers Cardio-Protection by Enhancing Mitochondrial Bioenergetics Through Coordination of Respiratory Complex and Supercomplex Assembly

www.nature.com/cr www.cell-research.com

ARTICLE OPEN NDUFAB1 confers cardio-protection by enhancing mitochondrial bioenergetics through coordination of respiratory complex and supercomplex assembly

Tingting Hou1, Rufeng Zhang1, Chongshu Jian1, Wanqiu Ding1, Yanru Wang1, Shukuan Ling2,QiMa1, Xinli Hu1, Heping Cheng1 and Xianhua Wang1

The impairment of mitochondrial bioenergetics, often coupled with exaggerated reactive oxygen species (ROS) production, is a fundamental disease mechanism in organs with a high demand for energy, including the heart. Building a more robust and safer cellular powerhouse holds the promise for protecting these organs in stressful conditions. Here, we demonstrate that NADH: ubiquinone oxidoreductase subunit AB1 (NDUFAB1), also known as mitochondrial acyl carrier protein, acts as a powerful cardio- protector by conferring greater capacity and efficiency of mitochondrial energy metabolism. In particular, NDUFAB1 not only serves as a complex I subunit, but also coordinates the assembly of respiratory complexes I, II, and III, and supercomplexes, through regulating iron-sulfur biosynthesis and complex I subunit stability. Cardiac-specific deletion of Ndufab1 in mice caused defective bioenergetics and elevated ROS levels, leading to progressive dilated cardiomyopathy and eventual heart failure and sudden death. Overexpression of Ndufab1 effectively enhanced mitochondrial bioenergetics while limiting ROS production and protected the heart against ischemia-reperfusion injury. Together, our findings identify that NDUFAB1 is a crucial regulator of mitochondrial energy and ROS metabolism through coordinating the assembly of respiratory complexes and supercomplexes, and thus provide a potential therapeutic target for the prevention and treatment of heart failure.

Cell Research (2019) 29:754–766; https://doi.org/10.1038/s41422-019-0208-x

INTRODUCTION sites in the complexes from oxidative damage, and stabilize Being so-called “powerhouses”, mitochondria comprise up to 40% individual complexes.11,15–21 Interestingly, SC formation is dyna- of the heart mass and are central to cardiac bioenergetics.1,2 As mically regulated in accordance with changes in cellular hazardous by-products of energy metabolism, reactive oxygen metabolism,22–25 and deficiency of SCs is associated with heart species (ROS) are also emitted from the respiratory electron failure.26 Therefore, enhancing SC formation might afford an transport chain (ETC), and excessive ROS emission causes effective therapeutic strategy to make the mitochondria more oxidative damage to proteins, lipids, and even genetic materials.3,4 efficient and safer powerhouses. Studies have shown that >50% of individuals with mutations in NADH:ubiquinone oxidoreductase subunit AB1 (NDUFAB1), also genes encoding mitochondrial proteins eventually develop known as mitochondrial acyl carrier protein,27 plays a multi- cardiomyopathy,5 suggesting that defective cardiac bioenergetics functional role in the organelle. It not only participates in the as well as its interlinked ROS metabolism is a fundamental disease synthesis of lipoic acid in the type II fatty acid biosynthetic mechanism in the heart.6,7 As such, building a more robust cellular pathway (FAS II),28,29 but also acts as an accessory subunit of powerhouse with decreased ROS emission holds the promise for complex I with a 2:1 stoichiometry.30 A recent study has shown protecting the heart in stressful conditions. that NDUFAB1 interacts with seven mitochondrial LYR motif- The most important player in mitochondrial energy production containing (LYRM) proteins: LYRM1, LYRM2, LYRM4/ISD11, LYRM5, is the ETC, which consists of four multi-heteromeric complexes LYRM7/MZM1L, SDHAF3/ACN9, and FMC1/C7orf55.31 Most of (complexes I–IV) in the inner membrane of the organelle. They them are linked to various mitochondrial metabolic processes, catalyze outward proton movement to build a transmembrane such as the assembly of complex II (for SDHAF3/ACN9) and electrochemical gradient that drives ATP synthase (complex V) for complex V (for FMC1/C7orf55), deflavination of the electron- ATP synthesis.8 Individual ETC complexes can also organize into transferring flavoprotein (for LYRM5), and biosynthesis of iron- supercomplexes (SCs) of different composition and stoichiometry, sulfur (FeS) centers (for LYRM4/ISD11).31 In yeast cells which lack which have recently been visualized by cryo-electron micro- complex I, Van et al. have shown that NDUFAB1 interacts with and scopy.9–14 Such SCs are thought to channel electron transfer stabilizes the FeS biogenesis complex to regulate the assembly of more efficiently, limit ROS production, protect vulnerable ETC complexes II and III and SCs.32 In addition, previous evidence in

1State Key Laboratory of Membrane Biology, Beijing Key Laboratory of Cardiometabolic Molecular Medicine, Peking-Tsinghua Center for Life Sciences, Institute of Molecular Medicine, Peking University, Beijing 100871, China and 2State Key Laboratory of Space Medicine Fundamentals and Application, China Astronaut Research and Training Center, Beijing 100094, China Correspondence: Heping Cheng ([email protected]) or Xianhua Wang ([email protected]) Received: 16 January 2019 Accepted: 2 July 2019 Published online: 31 July 2019

© The Author(s) 2019 Article 755 293T cells has also suggested a role of NDUFAB1 in complex I and the inter-ventricular septum was decreased whereas the left assembly and the existence of a complex I-independent mechan- ventricular diameter and volume were increased with age in cKO ism for Ndufab1 knockout-induced cell death.33 Thus, NDUFAB1 hearts (Supplementary information, Fig. S2b, c). These findings might be a core player in orchestrating mitochondrial energy indicate that cardiac-specific ablation of NDUFAB1 leads to dilated biogenesis. cardiomyopathy accompanied by cardiomyocyte hypertrophy, In this study, we aimed to test the potential protective role of interstitial fibrosis, and systolic and diastolic dysfunction, even- NDUFAB1 in the mammalian heart using cardiac-specific Ndufab1 tually resulting in heart failure and sudden death, revealing an knockout (cKO) and transgenic overexpression (TG) mouse essential role of NDUFAB1 in cardiac function. models. We find that NDUFAB1 coordinates the assembly of individual ETC complexes and SCs and acts as a crucial Mitochondrial dysfunctions in Ndufab1 cKO heart endogenous regulator of the efficiency and capacity of mitochon- To investigate the mechanism underlying the cardiomyopathy drial energy metabolism as well as ROS emission. More and heart failure induced by NDUFAB1 ablation, we next assessed importantly, NDUFAB1 exerts a cardio-protective effect when the changes of cardiac mitochondrial functions at different ages of heart is subjected to ischemia-reperfusion (IR) injury. Our findings cKO mice. Electron microscopic analysis revealed that cardiac substantiate that targeting mitochondria for greater bioenergetic mitochondria were in disarray along the sarcomeres with some capacity and repressed ROS emission with enhanced SC formation exhibiting irregularities in the cristae formation in 16-week-old is an effective strategy for cardio-protection, and mark NDUFAB1 cKO mice, whereas mitochondrial morphology was apparently as a potential therapeutic target for the prevention and treatment normal at 6 weeks, when the cardiac function hadn’t been of cardiomyopathy. impaired yet (Fig. 2a). The mitochondrial DNA content and cross- sectional area of individual mitochondria didn’t display detectable changes whereas the total mitochondrial volume fraction and RESULTS mitochondrial number were slightly increased in 16-week-old cKO Cardiac-specific ablation of NDUFAB1 causes progressive dilated mice (Supplementary information, Fig. S3). Measuring mitochon- cardiomyopathy leading to heart failure drial membrane potential (ΔΨm) in isolated cardiomyocytes with NDUFAB1 was widely expressed in different tissues and was the potential-sensitive fluorescent probe tetramethyl rhodamine particularly enriched in the heart (Supplementary information, methyl ester (TMRM) showed that NDUFAB1 ablation significantly Fig. S1a). Whole-body knockout of Ndufab1 was embryonic lethal decreased ΔΨm even at 6 weeks (Fig. 2b) when the heart

1234567890();,: (Supplementary information, Fig. S1b, c), showing that NDUFAB1 morphology and function were normal (Fig. 1), indicating is essential for developmental viability. To explore the potential mitochondrial dysfunction is an early event in cKO mice preceding cardiac functions of NDUFAB1, we generated a cardiac-specific the onset of cardiomyopathy. Measurements with the fluorescent knockout mouse model (cKO) wherein exon 3 of Ndufab1 was probe mitoSOX showed that mitochondrial ROS level was flanked by loxP sites (Ndufab1flox/flox, Supplementary information, significantly elevated in cKO cardiomyocytes compared to WT at Fig. S1b). The Ndufab1flox/flox mice were cross-bred with Mlc2v-Cre 10 weeks, and increased by 72% at 14–16 weeks, reaching a level mice to allow cardiomyocyte-specific deletion of Ndufab1,as comparable to that induced by 5 μg/mL antimycin A (Fig. 2c). The previously described.34 As shown in Fig. 1a and Supplementary cellular ATP level, as measured with the luciferin luminescence information, Fig. S1d, cardiac expression of Ndufab1 was assay, showed a trend of decrease in cKO mice from 6 to 8 weeks decreased by ~90% at the protein level in cKO cardiomyocytes old and was significantly lowered in 14–16-week or older cKO compared to wild-type (WT) cells. The cKO mice were smaller, had mice (Fig. 2d). This result suggests that, while the ATP level is a reduced body weight, and even started to lose weight initially tightly safeguarded in the heart,35–39 NDUFAB1 ablation precipitously after 14 weeks of age (Fig. 1b). Meanwhile, the curtails the energy reserve capacity, exhaustion of which impairs lifespan of cKO mice was markedly shortened, with sudden death ATP homeostasis. Taken together, these results indicate that beginning at ~12 weeks of age; the maximal lifespan was NDUFAB1 is essential for mitochondrial bioenergetics and ROS <19 weeks (Fig. 1c). metabolism. We next examined the effect of NDUFAB1 ablation on cardiac Next, we analyzed the transcriptome of left ventricles from cKO morphology and functions at different ages. The cKO hearts and WT mice by RNA sequencing (RNA-seq). Among 11494 genes progressively enlarged with age and developed dilated cardio- analyzed, 430 were downregulated and 1081 were upregulated myopathy (Fig. 1d; Supplementary information, Fig. S1e). Com- (Supplementary information, Tables S1 and S3). Interestingly, the pared to WT littermates, the heart weight/body weight ratio in top 5 Gene Ontology terms of the downregulated genes were cKO mice was increased by 12.8% at 6 weeks, 26.9% at 8 weeks, mainly associated with mitochondrial metabolism, including 51.5% at 12 weeks, and 241.7% at 16 weeks; and the heart weight/ metabolic process, fatty acid metabolism, lipid homeostasis, tibia length ratio was not significantly altered at 6 weeks and oxidation-reduction process (Fig. 2e; Supplementary information, increased by 45.5% at 8 weeks, 80.1% at 12 weeks, and 110.9% at Table S2), in good agreement with the bioenergetic defects in the 16 weeks (Fig. 1e). The slight increase of heart weight/body cKO heart (Fig. 2b–d). The downregulated genes were involved in weight ratio without change of heart weight/tibia ratio at 6 weeks the regulation of muscle contraction, consistent with the is due to mild reduction of body weight (Fig. 1b). At the cellular phenotype of cardiomyopathy (Fig. 2e). Meanwhile, the top level, the longitudinal-section area of cardiomyocytes was nearly enriched Gene Ontology terms of upregulated genes were mostly constant in WT but increased progressively in cKO mice, with no involved in immune system process, inflammatory response, cell significant change at 6 weeks and reaching a 78% increase at adhesion, and chemotaxis (Fig. 2e; Supplementary information, 16 weeks (Fig. 1f), indicative of cardiomyocyte hypertrophy. Table S4). We deduce that these upregulated processes might Masson staining showed that NDUFAB1 ablation induced exten- reflect compensatory as well as maladaptive responses to cardiac sive cardiac fibrosis early at 8 weeks (Fig. 1g; Supplementary damage due to mitochondrial dysfunctions. information, Fig. S1f). Consistent with the changes of cardiac morphology, echocardiography revealed that the ejection fraction Impaired assembly of mitochondrial ETC complexes and SCs in (EF) and fractional shortening (FS) of cKO hearts were unaltered at cKO hearts 6 weeks, but halved at 8 weeks; and the EF was diminished by To further understand the mechanism underlying these mito- 79% and FS by 84% at the age of 14 weeks (Fig. 1h), while the chondrial energy and ROS metabolism dysfunctions induced by heart rate remained unchanged (Supplementary information, NDUFAB1 ablation, we tested whether NDUFAB1 deficiency Fig. S2a). The thickness of both the left ventricular posterior wall affected lipoic acid synthesis in FAS II pathway. We measured

Cell Research (2019) 29:754 – 766 Article 756

Fig. 1 Cardiac-speciﬁc knockout of Ndufab1 (cKO) causes dilated cardiomyopathy leading to heart failure in mice. a Western blots of NDUFAB1 in wild-type (WT) and cKO cardiomyocytes. ATPB served as the loading control. b Left, photograph of WT and cKO mice at 14 weeks of age. Right, effect of cardiac NDUFAB1 ablation on mouse body weight (mean ± s.e.m.; n = 3–16 male mice per group per time point; **p < 0.01 versus WT at the same time point). c Kaplan–Meier survival curves for WT and cKO mice (n = 11 male mice per group; **p < 0.01 versus WT). d Representative longitudinal sections of hearts from male mice at 6–16 weeks (w) old. Scale bar, 5 mm. e Ratios of heart weight (HW) to body weight (BW) or tibia length (TL) (mean ± s.e.m.; n = 3–14 male mice per group; *p < 0.05, **p < 0.01 versus WT at the same age). f Progressive hypertrophy of cardiomyocytes from cKO mice. Left: representative confocal micrographs of cardiomyocytes stained with DCF (scale bar, 20 μm). Right: statistical analysis (mean ± s.e.m.; n = 97–611 cells from 4 male mice per group; ** p < 0.01 versus WT at the same age). g Fibrosis of left ventricular tissue (Masson’s trichrome staining; scale bar, 100 μm). h Echocardiographic analysis of cardiac function. EF, ejection fraction; FS, fractional shortening (mean ± s.e.m.; n = 3–9 male mice per group; **p < 0.01 versus WT at the same age)

pyruvate dehydrogenase activity as lipoic acid is its obligate using antibodies against NDUFB8 (complex I), SDHA (complex II), cofactor,40 and found similar activities in WT and cKO hearts UQCRFS1 or UQCRC1 (complex III), COXIV (complex IV), and ATPB (Supplementary information, Fig. S4a). Further, the protein (complex V), while complexes II and V mainly remained lipoylation status was not significantly altered in cKO hearts as individual entities in both WT and cKO groups (Fig. 3c). (Supplementary information, Fig. S4b). These results revealed Moreover, NDUFAB1 ablation decreased the SCs comprised of NDUFAB1 ablation did not affect FAS II pathway in the heart. complexes I, III, and IV in the hearts at both ages (Fig. 3c; Given the pivotal role of the mitochondrial ETC in energy Supplementary information, Fig. S6b). We also assessed the metabolism and ROS production, we next investigated whether abundance of individual complexes and found that complexes I, and how NDUFAB1 ablation impacted on the assembly and II, and III were significantly diminished in cKO mitochondria at activity of individual ETC complexes in the heart. First, we both ages (Fig. 3c; Supplementary information, Fig. S6b), con- measured the respiratory activity of isolated mitochondria from sistent with the aforementioned functional data (Fig. 3a). The level 6-week-old mice when the cardiac morphology and function were of individual complex IV was increased in cKO mitochondria due normal and 10-week-old mice when cardiomyopathy was already to decreased SC assembly (Fig. 3c; Supplementary information, developed. In the presence of substrates of either complex I, Fig. S6b). Further analysis showed the percentages of complex III- complex II or complex III, the oxygen consumption rates (OCRs) of and complex IV-containing SCs were decreased in cKO mitochon- cKO mitochondria were significantly decreased at both ages dria at both ages (Supplementary information, Fig. S6c). The (Fig. 3a; Supplementary information, Fig. S5a–c), whereas the OCR percentage of complex I-containing SCs was not significantly supported by complex IV substrate remained unchanged (Fig. 3a; changed because both the free and complexed complex I were Supplementary information, Fig. S5d). These data suggest that, proportionally diminished in cKO mitochondria (Supplementary rather than being merely a complex I subunit, NDUFAB1 alters the information, Fig. S6b, c). In-gel activity analysis in the presence of ETC function at multiple sites. complex I substrate revealed that the activity of either complex I Further, we assessed the assembly of ETC complexes using blue or SCs was markedly lowered in the absence of NDUFAB1 (Fig. 3b; native polyacrylamide gel electrophoresis (BN-PAGE). Consistent Supplementary information, Fig. S6a). Moreover, we measured the with previous reports,41,42 complexes I, III, and IV were assembled abundance of individual complexes and SCs in the hearts of into SCs, as evidenced by immunoblot analysis following BN-PAGE newborn mice (born within 24 h) and found that complexes I, II, III

Cell Research (2019) 29:754 – 766 Article 757

Fig. 2 Defective mitochondrial bioenergetics and ROS metabolism in Ndufab1 cKO mice. a Electron micrographs of mitochondria in left ventricular tissue from 6- and 16-week-old male mice (scale bars, 1 μm). b Decline of ΔΨm in cKO myocardium assessed by TMRM staining and confocal imaging of isolated cardiomyocytes. The intensity of FCCP-sensitive TMRM fluorescence was calculated and the intensity in cKO group was normalized to that of WT at the same age (n = 13–26 cells from three male mice per group; *p < 0.05, **p < 0.01 versus WT). c Progressive elevation of mitochondrial ROS measured with mitoSOX in isolated cardiomyocytes. Antimycin A (AA, 5 μg/mL) was used to induce massive ROS production. The intensities of mitoSOX fluorescence in the groups of cKO, WT + AA and cKO + AA were normalized to that of WT at the same age (n = 31–179 cells from 4–5 male mice per group; ** p < 0.01 versus WT). d ATP levels of isolated cardiomyocytes measured with luciferin assay (n = 3–4 male mice per group; ** p < 0.01 versus WT). e Top 5 enriched Gene Ontology terms of downregulated and upregulated genes. RNA-seq analysis was performed on cKO and WT hearts at 8–16 weeks (n = 4 male mice per group) and the genes upregulated or downregulated were analyzed and SCs were diminished while complex IV was not significantly whereas the other subunits examined were unaltered in the changed in cKO mitochondria (Supplementary information, absence of NDUFAB1 (Fig. 3e, f; Supplementary information, Fig. S7b, d). Likewise, in-gel activity analysis in the presence of Fig. S6e). This indicates that NDUFAB1 ablation selectively disrupts complex I substrate showed decreased activities of complex I and FeS-containing subunits of complexes II and III. Concomitantly, a SCs in cKO mitochondria of newborn mice (Supplementary complex III assembly intermediate lacking UQCRFS1, which is information, Fig. S7a, c). Meanwhile, knocking down Ndufab1 in incorporated into complex III at the last step,43 accumulated in cultured neonatal rat cardiomyocytes also decreased the abun- cKO mitochondria (Fig. 3c; Supplementary information, Fig. S7b). dance of complex I and SCs (Supplementary information, Fig. S8). These findings suggest that NDUFAB1 coordinates the assembly of All these results implicate that NDUFAB1 is essential for the ETC complexes and SCs through regulating FeS cluster biogenesis assembly of complexes I, II, III, and SCs in the heart. Furthermore, as reported previously in yeast cells lacking complex I.32 In this the early onset of impaired mitochondrial energetics induced by regard, we found that ISCU and NFS1, two important components disrupted assembly of ETC complexes and SCs indicates a of the FeS biogenesis complex,44,45 were down-regulated in causative role of this NDUFAB1 ablation-induced ETC defect in cKO cardiomyocytes (Supplementary information, Fig. S9). For heart failure. complex I, however, all subunits examined, regardless of whether By measuring the abundance of individual subunits of they contain (NDUFS1, NDUFS4, and NDUFS7) or don’t contain complexes I-III, we found that the FeS-containing subunits, SDHB (NDUFS6 and NDUFB8) FeS cluster(s), were dramatically decreased (succinate dehydrogenase subunit B) of complex II and UQCRFS1 in cKO mitochondria (Fig. 3d; Supplementary information, (ubiquinol-cytochrome c reductase, Rieske iron-sulfur Fig. S6d), indicating the involvement of additional mechanism polypeptide 1) of complex III, were both significantly decreased, by which NDUFAB1 regulates its assembly (see Discussion). Since

Cell Research (2019) 29:754 – 766 Article 758

Fig. 3 Defective respiration and impaired assembly of ETC complexes I-III and SCs in cKO hearts. a Effect of NDUFAB1 ablation on state III oxygen consumption rate (OCR) in isolated mitochondria at the age of 6 or 10 weeks. For each measurement, 100 μg mitochondria were used. Different respiratory substrates were used: malate/glutamate (Mala/glu, 5 mM) for complex I, succinate (Succ, 5 mM) for complex II, glycerol-3- phosphate (G3P, 5 mM) for complex III, and ascorbate (Asc, 2.5 mM)/N,N,N’,N’-tetramethyl-p-phenylenediamine (TMPD, 0.5 mM) for complex IV, and ADP (100 μM) (mean ± s.e.m.; n = 3–14 male mice of per group; *p < 0.05, **p < 0.01 versus WT at the same age). b In-gel activity of SCs and complex I at the age of 6 or 16 weeks. c BN-PAGE immunoblots of individual ETC complexes and SCs at the age of 6 or 16 weeks. SCs were visualized by antibodies against subunits of complex I (CI, NDUFB8), complex III (CIII, UQCRFS1 or UQCRC1), and complex IV (CIV, COX IV). Complex II was visualized by antibody against SDHA, and complex V with antibody against ATPB. d–f Western blots of subunits of complex I (d), complex II (e), and complex III (f) at the age of 6 or 16 weeks. Dagger represents FeS-containing subunits. ATP5A served as the loading control

the mRNA levels of the above subunits of complexes I–III were not decreased NAD+/NADH ratio and increased protein acetyla- significantly altered (Supplementary information, Fig. S10), the tion,46 however, we found that neither the NAD+/NADH ratio NDUFAB1-mediated regulation appears to occur at the protein nor protein acetylation levels were significantly altered in rather than the mRNA level, presumably by affecting protein Ndufab1 cKO cardiomyocytes (Supplementary information, stability. Therefore, NDUFAB1 coordinates the assembly of Fig. S11). These results suggest a specific role of NDUFAB1 in complexes I–III and SCs through its dual roles both as a regulator the assembly of respiratory complexes and SCs. In addition, of FeS biogenesis and as a complex I subunit. mitochondrial cardiolipin content was not altered in Ndufab1 Different from the effect of NDUFAB1 ablation on respiratory cKO mitochondria (Supplementary information, Fig. S12). All complexes, knockout of Ndufs4, another accessory subunit of these results substantiate that NDUFAB1 ablation-induced complex I, induced complex I deficiency but showed little effect cardiomyopathy is mainly caused by impaired assembly of ETC on other complexes.46 Meanwhile, NDUFS4 ablation led to complexes and SCs.

Cell Research (2019) 29:754 – 766 Article 759

Enhanced mitochondrial functions in Ndufab1 transgenic hearts NDUFAB1 might provide an effective strategy to build a more The results thus far reveal that NDUFAB1 plays an essential role in robust and efﬁcient powerhouse to beneﬁt the heart, providing coordinating the dynamic assembly of individual ETC complexes that endogenous NDUFAB1 is present at a sub-saturating level. To and SCs and thereby regulates mitochondrial bioenergetics and test this possibility, we generated an Ndufab1 transgenic mouse ROS metabolism. We reckoned that augmenting the abundance of model (TG) using a β-actin promoter (Supplementary information,

Cell Research (2019) 29:754 – 766 Article 760 Fig. 4 Enhanced mitochondrial bioenergetics and SC assembly in the heart of Ndufab1 transgenic mice (TG). TG and WT littermates of 2–4 months old were used. a Western blots for NDUFAB1 expression in WT and TG hearts. ATPB served as the loading control. b Decreased mitochondrial ROS level in TG cardiomyocytes. The intensity of mitoSOX ﬂuorescence of TG was normalized to that of WT (n = 34–41 cells from 3 male mice; **p < 0.01 versus WT). c Increased ΔΨm in TG cardiomyocytes. The intensity of FCCP-sensitive TMRM ﬂuorescence of TG was normalized to that of WT (n = 12–15 cells from three male mice per group; *p < 0.05 versus WT). d, e Whole-cell OCR measured with Seahorse using glucose/pyruvate (d) or palmitate (e) as the substrate. Arrows indicate the time of adding palmitate (PAL, 200 µM), oligomycin (Oligo, 1 µM), FCCP (0.5 µM), and rotenone/antimycin A (Rot/AA, 1 µM for each). f, g Statistical analysis of maximal OCR in d and e (mean ± s.e.m.; n = 10–13 wells from 3 male mice per group; *p < 0.05 versus WT). h Changes of respiratory control ratio (RCR) with different substrates in isolated mitochondria (mean ± s.e.m.; n = 3–8 male mice per group; *p < 0.05, **p < 0.01 versus WT). i In-gel activity of SCs and complex I. j Statistical analysis of i. The activity of TG was normalized to WT (mean ± s.e.m.; n = 5 male mice per group; *p < 0.05, **p < 0.01 versus WT). k BN-PAGE immunoblots of SCs and individual ETC complexes as in Fig. 3c. l Statistical analysis of k. The expression level of TG complexes and supercomplexes was normalized to that of WT (mean ± s.e.m.; n = 4–8 male mice per group; *p < 0.05 versus WT)

Fig. 5 NDUFAB1 overexpression protects the heart against IR injury. TG and WT male littermates of 2–4 months old were used. a Alleviated myocardial ROS production in TG hearts. Confocal images of DCF-stained myocardium after Langendorff perfusion and IR injury. Ischemia (IS) was mimicked by stopping perfusion. See Supplementary information, Fig. S19a for experimental protocol. Scale bars, 20 μm. b Statistical analysis of a (mean ± s.e.m.; n = 32–51 image frames from 4–7 male mice per group; **p < 0.01 versus WT, #p < 0.01 after IR versus basal). c Rotenone (Rot)-sensitive H2O2 production at complex I in WT and TG cardiac mitochondria (mean ± s.e.m.; n = 6 male mice per group; **p < 0.01 versus WT). Succinate (Succ, 5 mM) was used as the substrate of complex II and H2O2 was measured with Amplex red. d Images for simultaneous measurement of ΔΨm (TMRM, pseudocolor: red) and cardiomyocyte death (EBD uptake, pseudocolor: green) in Langendorff- perfused hearts subjected to IR injury. Note that all cells with intact ΔΨm are EBD-negative, and all EBD-positive cells display loss of ΔΨm, but not vice versa. Scale bars, 50 μm. e, f Percentages of cells with loss of ΔΨm (e) and EBD-positive cells (f) (mean ± s.e.m.; n = 9–10 male mice per group; *p < 0.05 versus WT, #p < 0.01 after IR versus basal). g Representative cross-sections of WT and TG hearts after IR injury. See Supplementary information, Fig. S19b for the experimental protocol. White, infarcted area (IF); red, rest of the area at risk (AAR); blue, tissue not at risk. The AAR was demarcated by dotted line. Scale bar, 1 mm. h, i Statistical analysis of AAR (h) and IF (i) reported as the ratios of AAR to left ventricular area (LV) and IF to AAR (n = 5–8 male mice per group; **p < 0.01 versus WT). j Measurement of serum LDH release after IR (n = 10–13 male mice per group; *p < 0.05 versus WT, #p < 0.01 after IR versus basal). k Echocardiographic analysis of cardiac function 48 h after IR injury. EF, ejection fraction; FS, fractional shortening (mean ± s.e.m.; n = 4 male mice per group; *p < 0.05 versus WT)

Fig. S13a) in which NDUFAB1 was overexpressed ~8-fold in the information, Fig. S13c, d). Cardiac morphology and functional heart (Fig. 4a; Supplementary information, Fig. S13b). Compared performance determined by echocardiography (Supplementary with WT littermates, TG mice grew normally and had similar information, Fig. S13e–g), mitochondrial morphology revealed by lifespans under our experimental conditions (Supplementary electron microscopy (Supplementary information, Fig. S13h–j),

Cell Research (2019) 29:754 – 766 Article 761 and mitochondrial DNA content (Supplementary information, (Supplementary information, Fig. S19b). In agreement with the Fig. S13l) were all comparable in WT and TG hearts. Nonetheless, ex vivo IR results, the area of infarction in TG was significantly the mitochondrial ROS level was markedly decreased in TG smaller than that in WT hearts (Fig. 5g–i). Meanwhile, lactate cardiomyocytes (Fig. 4b), suggesting an alleviation of basal dehydrogenase (LDH) release induced by IR injury was signifi- oxidative stress. The ΔΨm was significantly increased (Fig. 4c), cantly attenuated in TG mice (Fig. 5j). Echocardiography revealed and the maximal OCR with either glucose/pyruvate or palmitate as that TG mice exhibited significantly improved heart function at 48 substrate measured by Seahorse assay was higher in TG than in h after IR injury compared to WT mice (Fig. 5k). Altogether, these WT cardiomyocytes (Fig. 4d–g). These results imply an enhanced findings demonstrate that increasing the abundance of mitochon- reserve capacity for ATP production, although the homeostatic drial NDUFAB1 attenuates ROS production and protects the heart ATP level was unaltered (Supplementary information, Fig. S13k). against IR injury. The respiratory control ratio was significantly improved in the presence of substrates of complex I, II, or III in TG cardiac mitochondria (Fig. 4h), indicating greater efficiency of mitochon- DISCUSSION drial energy metabolism. Moreover, the electron flow assay The heart is quite sensitive to mitochondrial dysfunction such that showed enhanced activities of complexes I, II and III in TG most individuals with mutated mitochondrial proteins eventually mitochondria (Supplementary information, Fig. S14), indicating develop cardiomyopathy.5 Not only are mitochondria the that NDUFAB1 overexpression augments the ETC activity. predominant powerhouses of the heart, supplying more than At the molecular level, in-gel activity analysis showed that 90% of the ATP,2 but also excessive mitochondrial ROS are the activity of both complex I and SCs were significantly enhanced in culprits in many oxidative heart diseases including IR injury. Here TG hearts (Fig. 4i, j), and BN-PAGE analysis showed that NDUFAB1 we have shown, for the first time, that NDUFAB1 is a key overexpression significantly increased the contents of complex I endogenous regulator of mitochondrial bioenergetics and ROS and SCs (Fig. 4k, l; Supplementary information, Fig. S15a), whereas metabolism in the mammalian heart and a cardio-protector when the expression of all subunits examined, including those for the heart is subjected to stress and injury. Our ex vivo and in vivo complexes I-III, was unchanged (Supplementary information, experimental results have established the essential role of Fig. S15b, c). Importantly, overexpression of NDUFAB1 in cultured NDUFAB1 in the maintenance of normal cardiac functions and neonatal cardiomyocytes enhanced the abundance of complex I the cardio-protective effects of NDUFAB1 overexpression. In the and SCs without significant effect on the other complexes cKO heart, NDUFAB1 ablation led to lowered ΔΨm and diminished (Supplementary information, Fig. S16). In addition, NDUFAB1 ATP production. Meanwhile, ROS emission was elevated to overexpression did not impact on FAS II pathway in the heart, pathologically high levels, aggravating the situation of insufficient because neither pyruvate dehydrogenase activity nor protein energy supply. Cardiac performance deteriorates precipitously lipoylation status was significantly altered in TG mitochondria once the energy reserve is exhausted and the ATP level can no (Supplementary information, Fig. S17). The NAD+/NADH ratio and longer be held constant. Decompensatory myocardial remodeling protein acetylation levels were also comparable in the WT and TG also ensues, manifested as myocardial fibrosis, cardiomyocyte hearts (Supplementary information, Fig. S18). Taken together, hypertrophy, and dilation of the ventricles with thinning of the NDUFAB1 overexpression enhances the abundance of complex I ventricular walls, rapidly leading to heart failure and sudden and SCs, conferring on mitochondria greater capacity and death. Conversely and more importantly, increasing NDUFAB1 efficiency of energy metabolism and less ROS emission. abundance makes the mitochondria not only more robust, featured with augmented energy reserve capacity and efficiency, NDUFAB1 overexpression protects the heart against IR injury but also safer in terms of repressed ROS emission. The cardiac performance as well as the lifespan of TG mice was Both the improved bioenergetics and the diminished ROS similar to that of WT littermates under unchallenged conditions production appear to be indispensable for NDUFAB1-mediated (Supplementary information, Fig. S13). However, TG mice might be cardio-protection. On one hand, as the heart is highly demanding more resistant to cardiac injury because of augmented capacity of energy, increasing the energy reserve capacity would make the and efficiency of mitochondrial bioenergetics along with attenu- heart more resistant to insults that tend to curtail ATP production. ated ROS production. Next, we subjected the WT and TG hearts to On the other hand, since excessive ROS constitute a crucial driving IR injury to unmask this potential cardio-protective role of factor for IR injury in the heart,47–49 repression of basal ROS, and NDUFAB1. First, we applied an ex vivo IR protocol based on particularly the ROS burst during IR, should be beneficial. Langendorff perfusion with 30-min ischemia followed by 30-min Specifically, we have found that NDUFAB1 overexpression reperfusion (Supplementary information, Fig. S19a). Since exces- represses ROS production at the onset of reperfusion after sive ROS production during IR is a major contributor to IR ischemia. In agreement, our unpublished data showed that injury,47–49 we tracked the ROS production with indicator NDUFAB1 was decreased in human failing hearts and in mouse 5-(and-6)-chloromethyl-2′,7′-dichlorodihydrofluorescein diacetate hearts after IR, supporting the possibility that NDUFAB1 desregu- acetyl ester (DCF) (Fig. 5a). In WT hearts, there was a trend of lation may contribute to mitochondrial dysfunction observed in increasing ROS during ischemia followed by a prominent ROS heart failure and IR patients. Altogether, our results substantiate burst immediately after reperfusion (Fig. 5a, b). Remarkably, the the emerging concept that energetically more robust and ROS- ROS burst after IR was ameliorated by 29% in TG hearts (Fig. 5a, b). repressed mitochondria can protect the heart in stressful Furthermore, we measured H2O2 production in the presence of situations. succinate in isolated cardiac mitochondria using the Amplex Red Enhanced assembly of respiratory SCs induced by NDUFAB1 50 assay and found that the rotenone-sensitive H2O2 production overexpression contributes to improved bioenergetic capacity and was significantly attenuated in TG mitochondria (Fig. 5c). efficiency reflected by increased mitochondrial membrane poten- Because loss of ΔΨm not only manifests a defective bioener- tial and respiration and repressed ROS production. Formation of getics but also is an early event of cell death,51,52 we further SCs has been shown to enhance the catalytic activity of individual 53 assessed cardiac damage by monitoring loss of ΔΨm with TMRM components, facilitate electron transfer, and reduce ROS and membrane integrity with evans blue dye (EBD). Our results production.17,22,41 It has been recently reported that accumulated showed that cells that both lost TMRM and were EBD-positive succinate during ischemia fuels ROS burst at complex I through were significantly fewer in TG than WT hearts (Fig. 5d–f). reverse electron transport.54 In the TG hearts, however, even Moreover, we adopted an in vivo IR experimental protocol in though the complex I content was increased, both reverse which a 30-min ischemia was followed by 24-h reperfusion electron transport and succinate accumulation might be inhibited

Cell Research (2019) 29:754 – 766 Article 762 by enhanced SC formation and heightened respiration, decreasing Peking University accredited by AAALAC International (IMM- IR-induced ROS production. Conversely, NDUFAB1 ablation caused ChengHP-14). impaired SC formation, leading to deficiency of mitochondrial respiration and individual ETC complex activity. Elevated mito- Generation of Ndufab1 knockout and transgenic mice chondrial ROS levels were also observed in cKO cardiomyocytes. Floxed Ndufab1 mice were generated by standard techniques Thus, our findings implicate that enhancing SC formation can be using a targeting vector containing a neomycin-resistance an effective strategy to enhance mitochondrial bioenergetics and cassette flanked by FRT sites. Briefly, exon 3 of the Ndufab1 gene limit ROS production. was inserted into two flanking LoxP sites. After electroporation of NDUFAB1 coordinates the assembly of ETC complexes and SCs, the targeting vector into embryonic stem (ES) cells v6.5 (129 × mainly by regulating the biogenesis of FeS clusters as reported C57), G418-resistant ES cells were screened for homologous previously.32 In the absence of NDUFAB1, the assembly of recombination by Southern blot. Two heterozygous recombinant complexes I-III, but not complex IV, was severely impaired. This ES clones were identified and microinjected into blastocysts from result is in good agreement with the fact that complexes I-III all C57BL/6 J mice to generate germline-transmitted floxed hetero- contain FeS clusters, but complex IV does not.8 Furthermore, only zygous mice (Ndufab1f/+). Homozygous Ndufab1-floxed mice the FeS-containing subunits of complexes II and III were selectively (Ndufab1f/f) were obtained by inbreeding the Ndufab1f/+ mice. down-regulated, and the premature complex III lacking the FeS- Cardiac-specific Ndufab1 knockout mice were generated as containing subunit UQCRFS1 accumulated in cKO mitochondria. previously described.34 Briefly, Ndufab1f/f mice were bred with For complex I, however, an additional mechanism must be Mlc2v-Cre mice in which Cre recombinase expression was invoked because all subunits examined, including those lacking controlled by the myosin light chain 2v promoter to generate FeS clusters, were downregulated in the absence of NDUFAB1. double heterozygous Mlc2v-Cre and Ndufab1 floxed mice (Mlc2v- When NDUFAB1 was upregulated, greater effects were found on Cre/ Ndufab1f/+). The mice were then backcrossed with homo- the assembly of complex I than complex II or III. These qualitative zygous Ndufab1f/f mice to generate Mlc2v-Cre+/Ndufab1f/f cardiac- and quantitative differences are not unexpected, because cryo- specific Ndufab1 knockout mice and Mlc2v-Cre−/Ndufab1f/f electron microscopy has identified two copies of NDUFAB1 in littermate controls. Mice were genotyped by PCR analysis using complex I, and these are critically positioned for inter-subunit tail DNA, and flox primers (forward, ACAAATTCTCCCTGATGTCCTT; interactions within the complex.55 Therefore, NDUFAB1 upregula- reverse, TGTTCAACTTCATTTTGAGGTGGT) and Cre primers (for- tion enhances complex I assembly presumably through stabilizing ward, GCGGTCTGGCAGTAAAAACTATC; reverse, GTGAAA CAGCAT inter-subunit interactions. The enhanced complex I then promotes TGCTGTCACTT) were used. SC formation in TG mitochondria, in agreement with the proposed To generate global Ndufab1 knockout mice, Ndufab1f/f mice model that complex I assembly is completed before SC were bred with Prm-Cre mice in which Cre recombinase formation.56,57 Altogether, it is possible that NDUFAB1 modulates expression was controlled by the mouse protamine 1 promoter SC formation indirectly by regulating the assembly of individual in the male germ line to generate heterozygous Prm-Cre and complexes I-III, because it has been suggested that SCs form after Ndufab1 floxed mice (Prm-Cre+/Ndufab1f/+). The male mice were complete assembly of the individual complexes.11,33,56 Meanwhile, bred with wild-type (WT) C57BL/6 J females to generate hetero- our data also suggest the alternative possibility that NDUFAB1 zygous Ndufab1+/− mice. The heterozygous mice were inter- participates directly in the assembly of SCs. If this is the case, crossed to generate homozygous Ndufab1−/− mice. Mice were NDUFAB1 might reversely regulate the abundance of individual genotyped by PCR analysis using tail DNA, and WT primers complexes by affecting SC formation, since SCs have been (forward, ACAAATTCTCCCTGATGTCCTT; reverse, ACTGTATTCATG suggested to stabilize individual complexes.16,19–21 CCAGCATGTG) and KO primers (forward, ACAAATTCTCCCTGATGT In summary, we demonstrate that NDUFAB1 plays an essential CCTT; reverse, TCTGTTGTGCCCAGTCATAG) were used. role in mitochondrial bioenergetics and cardiac function To generate pan-tissue transgenic mice expressing Ndufab1, the through coordinating the assembly of respiratory complexes Ndufab1 cDNA was cloned into the pUCCAGGS vector down- and SCs. Mechanistically, these effects are mediated mainly by stream of the chicken β-promoter. The construct was linearized its dual roles both as a regulator of FeS biogenesis and as a with HindIII and PvuI to release the transgenic cassette, purified complex I subunit. Augmenting NDUFAB1 abundance is with a DNA purification kit (Qiagen), and microinjected into sufficient to enhance the energy metabolism efficiency and fertilized eggs of C57BL/6J mice. The mice were genotyped by PCR reserve capacity of mitochondria, and to lower ROS production. analysis using the primers AGCCTCTGCTAACCATGTTC (forward) It also protects the heart against injurious insults. Given recent and GTCCAAACTGTCTAAGCCCA (reverse). cryo-EM evidence for respiratory SCs,9–14 the present study The mice were housed under a 12-h-light cycle; food and water underscores the dynamic regulation of SC formation and reveals were provided ad libitum. the possibility to target SC formation as a means to build stronger and safer powerhouse and hence stronger heart. Echocardiography Overall, we identify a novel target to build a more robust Mice were anesthetized with isoflurane (2% in 100% O2 at mitochondria powerhouse for a stronger heart. It is thus 0.5 L/min) or avertin (2.5%, 0.01 mL/g, intraperitoneally, i.p.) and tempting to speculate that targeting NDUFAB1 to promote transthoracic echocardiography was performed using a VEVO- ETCcomplexandSCassemblymightaffordapotential 2100 Imaging System (Visual Sonics). The heart was imaged using therapeutic strategy for the prevention and treatment of M-mode and two-dimensional measurements were made at the mitochondrial bioenergetics-centered heart diseases. level of the papillary muscles. Data represent the average of at least five separate scans in a random-blind fashion. Left ventricular posterior wall thickness at the end-diastolic and end-systolic MATERIALS AND METHODS phases, left ventricular internal diameter in end-diastole and end- Study approval systole, and end-diastolic and end-systolic interventricular septum All animal experiments were carried out following the rules thickness were measured from the M-mode image. FS and EF were of the American Association for the Accreditation of Laboratory used to assess systolic function. Animal Care International and the Guide for the Care and Use of Laboratory Animals published by the US National Masson trichrome staining Institutes of Health (NIH Publication No. 85-23, revised 1996). Mouse hearts were excised and fixed in 4% paraformaldehyde All procedures were approved by the Animal Care Committee of overnight, embedded in paraffin, and sectioned at 4 μm. Fibrosis

Cell Research (2019) 29:754 – 766 Article 763 was visualized by Masson trichrome staining. Ten fields were protein concentration of the mitochondrial preparation was chosen randomly and imaged under an Olympus BX51 light determined with a NanoDrop Microvolume Spectrophotometer. microscope at ×20 magnification. ImageJ was used for fibrosis Mitochondrial respiration was evaluated by measuring oxygen quantification. consumption with a Clark-type oxygen electrode (Strathkelvin 782 2-Channel Oxygen System version 1.0; Strathkelvin Instruments, Transmission electron microscopy Motherwell, UK). Briefly, isolated mitochondria were re-suspended Freshly excised hearts were perfused with 1% glutaraldehyde and in respiration buffer (in mM: 225 mannitol, 75 sucrose, 10 KCl, 10 3 4% paraformaldehyde for 30 min. After dissecting into 1–2mm Tris-HCl, 5 KH2PO4, pH 7.2) at 25 °C with different substrates. State blocks, samples were immediately fixed with 2.5% glutaraldehyde II and III OCRs were each measured in the absence and presence and 4% paraformaldehyde, then post-fixed with 1% osmium of 100 μM ADP. The respiratory control ratio was defined as the tetroxide. After dehydration in a graded series of acetone, the ratio of the state III to the state II respiratory rate. samples were embedded in Spurr resin and sectioned with a glass knife on a Leica Ultracut R cutter. Ultra-thin sections (70–90 nm) BN-PAGE analysis of mitochondrial SCs and in-gel activity were imaged with a transmission electron microscope (Tecnai G2 BN-PAGE was conducted using the NativePAGETM system (Invitro- 20 Twin, FEI). Mitochondrial area was calculated by analyzing gen). Briefly, isolated cardiac mitochondria were solubilized by images of mitochondria with a clear outer membrane on the digitonin (4 g/g protein) for 15 min on ice, then centrifuged at micrographs using ImageJ freehand selection, and the number of 15,000 rpm at 4 °C for 30 min. After centrifugation, the super- pixels was converted to the actual mitochondrial cross-sectional natants were collected and the protein concentration was area according to the magnification and pixel size of each determined by BCA analysis (ThermoFisher). Coomassie blue G- micrograph. Mitochondrial volume fraction was analyzed by point- 250 (Invitrogen) was added to the supernatant to obtain a dye/ counting method.58 Briefly, a square lattice grid of 400 intersec- detergent mass ratio of 4/1 and then the protein was loaded into a tions was placed over each micrograph and the number of times a 4–16% non-denaturing polyacrylamide gel (Invitrogen). After mitochondrion located at an intersection was counted. The final electrophoresis, proteins were transferred to a PVDF membrane volume fraction was calculated by dividing the number with the and then probed with specific antibodies against subunits of total intersection number. complex I (NDUFB8), complex II (SDHA), complex III (UQCRFS1 and UQCRC1), complex IV (COX IV), and complex V (ATPB or ATP5A). Isolation of adult mouse cardiomyocytes Blots were visualized using secondary antibodies conjugated with Single ventricular myocytes were enzymatically isolated from the IRDye (LI-COR, Lincoln, NE, USA) and an Odyssey imaging system hearts of WT, Ndufab1 cardiac-specific knockout, or pan-tissue (LI-COR). The anti-NDUFB8, anti-UQCRFS1, or anti-COXIV immuno- transgenic mice, as described previously.59,60 Freshly isolated blot bands with high molecular weight were used to indicate SCs cardiomyocytes were plated on laminin-coated (Sigma) culture containing complex I, complex III, or complex IV. dishes for 1 h and then the attached cells were maintained in For in-gel activity analysis, 2.5 mg/mL nitro blue tetrazolium and Dulbecco’s modified Eagle’s medium (Invitrogen, Carlsbad, CA, 0.5 mg/mL NADH in 2 mM Tris-HCl (pH 7.4) were added to the USA) along with 10% FBS (Hyclone), 5 mM BDM (Sigma), and 1% 4–16% non-denaturing polyacrylamide gel after electrophoresis insulin-transferrin-selenium supplement (Invitrogen) until use. and incubated for 15 min at 37 °C. The reaction was stopped with 10% acetic acid. The activity was determined by analyzing the Isolation and culture of neonatal cardiomyocytes color development using the Odyssey imaging system (LI-COR). Ventricular myocytes were isolated from 1-day-old Sprague- Dawley rats, as described previously.61 Myocytes were plated at Whole-cell respiration analysis 1.5 × 105 cells/cm2 in DMEM (Invitrogen) supplemented with 10% Mitochondrial respiration was measured using the Seahorse XF24 FBS (Gibco) in the presence of 0.1 mM 5-bromo-2-deoxyuridine Extracellular Flux Analyzer (Seahorse Bioscience, North Billerica, (Sigma). Adenovirus infection or siRNA transfection was imple- MA, USA) according to the manufacturer’s instructions. Briefly, mented after 24 h quiescence in serum-free DMEM following isolated cardiomyocytes were seeded onto an XF24 microplate at 48–72 h culture in DMEM containing 10% FBS. For adenovirus 300 cells/well. The cellular OCR was monitored in unbuffered assay infection, cells were infected with adenovirus carrying LacZ or the medium (Sigma D5030) supplemented with (in mM) 2 GlutaMAX Ndufab1 gene at an m.o.i. of 20. For siRNA transfection, 100 nM (Gibco), 2.5 sodium pyruvate, and 25 glucose (pH 7.4 at 37 °C), siRNA was transiently transfected using Lipofectamine RNAiMax following the sequential addition of oligomycin (1 μM), carbonyl (Invitrogen) according to the manufacturer’s instructions. The cyanide 4-(trifluoromethoxy) phenylhydrazone (FCCP, 500 nM), knockdown efficiency was assessed by western blot. The double- and rotenone (1 μM) and antimycin A (1 μM) or in KHB buffer stranded RNA sequences for knockdown of rat Ndufab1 gene and (in mM, 111.3 NaCl, 4.7 KCl, 2.0 MgSO4, 1.2 Na2HPO4, 2.5 glucose, the negative control (NC) are listed in the following table: 0.5 L-Carnitine). For fatty acid supported respiration, the 2.5 mM sodium pyruvate and 25 mM glucose were replaced with 200 μM palmitate. Maximal OCR was calculated by subtracting the OCR in siRNA name Sense Anti-sense the presence of rotenone and antimycin A from that in the siNdufab1-1 CACUGACGUUAGAAGGAAUTT AUUCCUUCUAACGUCAGUGTT presence of FCCP. siNdufab1-2 GCUCUCAGUAAAUUCUCAUTT AUGAGAAUUUACUGAGAGCTT fl NC UUCUCCGAACGUGUCACGUTT ACGUGACACGUUCGGAGAATT Electron ow assay The electron flow in isolated mitochondria was measured using the Seahorse XF24 Extracellular Flux Analyzer (Seahorse Bioscience, North Billerica, MA, USA) according to the manufacturer’s instructions. Briefly, 2.5 μg of mouse heart Isolation and respiration measurement of cardiac mitochondria mitochondria were plated in each well of the XF24 plate Mouse hearts were washed with ice-cold isolation buffer (210 mM containing 1 × MAS (in mM: 70 sucrose, 220 mannitol, 5 mannitol, 70 mM sucrose, 5 mM HEPES (pH 7.4), 1 mM EGTA, and KH2PO4,5MgCl2, 2 HEPES, 1 EGTA, 0.2% BSA, pH 7.4 at 37 °C) 1 mg/mL BSA), minced, and homogenized. After the homogenate supplemented with 10 mM pyruvate, 2 mM malate and 4 μM was centrifuged at 4 °C for 10 min at 700 × g, the supernatant was FCCP. The mitochondrial OCR was monitored following the collected and further centrifuged at 4 °C for 10 min at 12000 × g. sequential addition of rotenone (2 μM), succinate (10 mM), The pellet was re-suspended for functional assessment. The antimycin A (4 μM) and ascorbate/TMPD (10 mM/100 μM). The

Cell Research (2019) 29:754 – 766 Article 764 electron flow of different complexes was revealed by their previously.51,52 Ischemia (30 min) was achieved by clamping the inhibitor-sensitive OCR. perfusion line and reperfusion (30 min) by releasing the clamp. Confocal images were captured from multiple, randomly selected Measurement of H2O2 production regions prior to, during ischemia, and after the IR, at excitation H2O2 production in isolated mitochondria was measured using wavelengths of 488, 543, and 633 nm and emission wavelengths the Amplex® Red hydrogen peroxide kit (Molecular Probes). of 500–540, 565–595, and >650 nm for DCF, TMRM, and EBD, Briefly, isolated mitochondria were incubated in an assay respectively. medium with 10 mM succinate, 50 mM Amplex red, and 5 units/mL HRP in each well of a 96-well plate. The increase in Confocal microscopy Amplex red fluorescence was followed over 60 min at room An inverted confocal microscope (Zeiss LSM 710) with a 40×, temperature with excitation at 530 nm and emission at >590 1.3 NA oil-immersion objective was used for imaging. For nm in a 96-well microplate reader (Biotek, Winooski, VT, USA). mitoSOX measurement, the indicator (5 μM) was loaded into H2O2 production was indexed by the increase of Amplex red isolated cardiomyocytes at 37 °C for 30 min followed by three fluorescence per min. washes with Tyrode’s solution. The mitoSOX fluorescence was measured by excitation at 514 nm and emission at 580–740 ATP measurement in isolated cardiomyocytes nm. To measure mitochondrial membrane potential in isolated The ATP was extracted from fresh isolated mouse cardiomyocytes cardiomyocytes, TMRM (50 nM) was loaded at 37 °C for 10 min with cold 2.5% trichloroacetic acid. The cellular ATP content was followed by three washes and the fluorescence was measured measured by luciferase assay according to manufacturer’s by excitation at 543 nm and emission at >560 nm. The intensity instructions (Promega, Madison, WI, USA). of FCCP (5 μM)-sensitive TMRM fluorescence was calculated and used to assess mitochondrial membrane potential. NAD+, NADH, and cardiolipin measurement For NAD+ and NADH measurement, 20 mg heart tissue was RNA-seq analysis homogenized and measured with a commercially available kit Using TRIzol reagent (Invitrogen), total RNA was isolated from according to the manufacturer’s instructions (BioAssay). For left ventricular myocardium in cKO and WT male mice cardiolipin measurement, 100 μg mitochondria were lysed by (8–16 weeks old) according to the manufacture’s protocol. freeze thawing and measured with a commercially available kit The quality of the extracted RNA was controlled using Agilent according to the manufacturer’s instructions (BioVision). 2100. A sequencing library was prepared using the NEBNex- t®UltraTM RNA Library Prep Kit for Illumina® (New England In vivo and ex vivo myocardial IR models Biolabs, Ipswich, MA, USA) following the manufacturer’s For in vivo IR surgery, 2–4-month-old male mice were anesthe- recommendations. Deep sequencing was then performed on tized with avertin (2.5%, 0.01 mL/g, i.p.) and ventilated via a an Illumina Hiseq 4000 platform and 150 bp paired-end reads tracheostomy on a Minivent respirator. A midline sternotomy was were generated. To obtain the differentially expressed genes, performed, and a reversible coronary artery snare occluder was RNA-seq reads were first aligned to the mouse genome placed around the left anterior descending coronary artery. (genome version mm9) with tophat2 (version 2.1.1; main Myocardial IR was induced by tightening the snare for 30 min parameters: –read-mismatches 8, –min-anchor 8, –segment- and then loosening it for 24 h. To measure the infarct size, the length 30, –segment-mismatches 2) and all the mapping animals were anesthetized with avertin (2.5%, 0.015 mL/g, i.p.) and results were evaluated as previously described.62 DEseq2 heparinized (400 USP U/kg, i.p.) at the end-point. The heart was packages were used to obtain differentially expressed genes excised and the ascending aorta was cannulated (distal to the withacutoffoffold-changeinparameP-value < 0.05. The sinus of Valsalva), then perfused retrogradely with Alcian blue dye functional annotation and pathway enrichment analysis for (1%) to visualize the area not at risk. The coronary artery was re- both upregulated and downregulated genes were performed occluded at the site of occlusion before perfusion with Alcian blue. using Database for Annotation, Visualization, and Integrated Then, the heart was frozen at −80 °C for 10 min and cut into 1 mm Discovery63,64 with a P-value cutoff of 0.05 under the slices (5–6 slices per heart), which were then incubated in sodium Benjamini test. phosphate buffer containing 1.5% 2,3,5-triphenyl-tetrazolium chloride for 15 min to visualize the unstained infarcted region. Western blot analysis Infarct area, area at risk (AAR), and left ventricular area were Isolated cardiomyocytes were homogenized with denaturing determined by planimetry using ImageJ. The infarct size was lysis buffer, and 10–50 μg of protein per sample was separated calculated as infarct area divided by AAR. For LDH measurement, on 12% SDS-PAGE. After electrophoresis, proteins were blood samples were collected before surgery and after 24 h of transferred to a PVDF membrane and then probed with specific reperfusion, and centrifuged for 10 min at 3000 rpm to obtain antibodies. Blots were visualized using secondary antibodies serum. LDH was spectrophotometrically assayed using a kit from conjugated with IRDye (LI-COR) and an Odyssey imaging Jingyuan Biology (Shanghai, China). system (LI-COR). The antibodies for NDUFAB1, SDHB, and For ex vivo IR, the heart was excised from male mice SDHD were from Origene, those for NDUFB8, NDUFS4, SDHA, (2–4 months old), and the ascending aorta was cannulated with UQCRC1, UQCRB, ATPB, ATP5A, AcK103, and lipoic acid from a customized needle. The heart was perfused in the Langendorff Abcam, and those for NDUFS1, NDUFS6, NDUFS7, UQCRFS1, configuration under constant perfusion pressure (1000 mm H2O) COXIV, ISCU, and NFS1 from ProteinTech. with oxygenated (100% O2) Tyrode’s solution (pH 7.4 at 37 °C). The Langendorff-perfused heart was placed on the stage of the RNA isolation and real-time PCR (RT-PCR) confocal microscope and images were captured ~30 μm deep into Total RNA was isolated from freshly isolated mouse cardiomyo- the epimyocardium of the left ventricle. Motion artifacts due to cytes with TRIzol reagent (Invitrogen) according to the manufac- spontaneous beating of the heart (~480/min) were minimized turer’s instructions, and then converted to cDNA using TransScript using 10 µM blebbistatin (Sigma). After a 10-min stabilization One-Step gDNA Removal and cDNA Synthesis Mix (Transgen period, the heart was perfused with 15 μM DCF (an indicator of Biotech). Quantitative RT-PCR reactions were performed using ROS) or 500 nM TMRM (an indicator of mitochondrial membrane Trans Start Green qPCR Super Mix (TransGen Biotech) and a potential) and 0.05% EBD (an indicator of plasma membrane BioRad CFX96 Touch RT-PCR Detection System. The primers used integrity) for 20 min. The IR protocol used was as described were as follows:

Cell Research (2019) 29:754 – 766 Article 765 10. Sousa, J. S., Mills, D. J., Vonck, J. & Kuhlbrandt, W. Functional asymmetry and Gene Forward Reverse electron flow in the bovine respirasome. Elife 5, https://doi.org/10.7554/ Ndufab1 CTGAGGGAATCCGGAGGAGA GTCTAACGTCAGTGGGGGTG eLife.21290 (2016). Ndufs1 CTCCTCTTGCCCTTGACTGG CCAGCCCTTCATTACAGGCA 11. Letts, J. A., Fiedorczuk, K. & Sazanov, L. A. The architecture of respiratory super- Ndufs4 ATCCACTTGGAAGCTGGCAGACAA AGACTGCATGTTATTGCGAGCAGG complexes. Nature 537, 644–648 (2016). 12. Gu, J. et al. The architecture of the mammalian respirasome. Nature 537, 639–643 Ndufs6 CCTGGAGCGATTCTGGGATAACTT TGGGCTTCGAGCTAACAATGGTGT (2016). Ndufs7 CAGAGTTCATCAGAGTGTAGCC CGGGGAAGATGAGAGAGCTTG 13. Guo, R. et al. Architecture of human mitochondrial respiratory megacomplex Ndufb8 CATCTCTTCGGCTTTGTGGC CAAAAAGCCCATCAAGCCTCC I2III2IV2. Cell 170, 1247–1257 (2017). Sdha AGATTGTGCCTGGTCTGTATGCCT AGGCTCTGCCAAAGACTACAAGGT 14. Gong, H. et al. An electron transfer path connects subunits of a mycobacterial – Sdhb ACCCCTTCTCTGTCTACCG AATGCTCGCTTCTCCTTGTAG respiratory supercomplex. Science 362, 6418 6428 (2018). 15. Porras, C. A. & Bai, Y. Respiratory supercomplexes: plasticity and implications. Sdhd TGGTACCCAGCACATTCACC GGGTGTCCCCATGAACGTAG Front. Biosci. 20, 621–634 (2015). Uqcrc1 GACAACGTGACCCTCCAAGT AGGATGTTTTTGCCCCGAGT 16. Milenkovic, D., Blaza, J. N., Larsson, N. G. & Hirst, J. The enigma of the respiratory Uqcrc2 GGCTTGTTCGTTAAAGCAGGCAGT TGCCTTCTACAGTGTACGCCATGT chain supercomplex. Cell Metab. 25, 765–776 (2017). Uqcrfs1 TCCCTGAAGGGAAGAACATGGCT TGCAGCTTCCTGGTCAATCTCCTT 17. Maranzana, E. et al. Mitochondrial respiratory supercomplex association limits production of reactive oxygen species from complex I. Antioxid. Redox Signal. 19, Gapdh ACCACAGTCCATGCCATCAC TCCACCCACCCTGTTGCTGTA 1469–1480 (2013). 18. Lopez-Fabuel, I. et al. Complex I assembly into supercomplexes determines dif- ferential mitochondrial ROS production in neurons and astrocytes. Proc. Natl Acad. Sci. USA 113, 13063–13068 (2016). Statistics 19. Schagger, H. et al. Significance of respirasomes for the assembly/stability of Data are expressed as mean ± s.e.m. When appropriate, Student’s human respiratory chain complex I. J. Biol. Chem. 279, 36349–36353 (2004). t-test was applied to determine statistical significance. The log- 20. Acin-Perez, R. et al. Respiratory complex III is required to maintain complex I in – rank test was used for survival curves. P < 0.05 was considered mammalian mitochondria. Mol. cell 13, 805 815 (2004). statistically significant. 21. Diaz, F., Fukui, H., Garcia, S. & Moraes, C. T. Cytochrome c oxidase is required for the assembly/stability of respiratory complex I in mouse fibroblasts. Mol. Cell. Biol. 26, 4872–4881 (2006). 22. Lapuente-Brun, E. et al. Supercomplex assembly determines electron flux in the ACKNOWLEDGEMENTS mitochondrial electron transport chain. Science 340, 1567–1570 (2013). We thank Ms. Yuli Liu for her kind help in echocardiogram test, Dr. Lin Pan and Ms. 23. Guaras, A. et al. The CoQH2/CoQ ratio serves as a sensor of respiratory chain Dongwei Ma for their help in histological analysis, Drs. Liping Wei and Chuanyun Li efficiency. Cell Rep. 15, 197–209 (2016). for their help in bioinformatics analysis, Dr. Rongli Zhang and Ms. Saifang Yan for 24. Acin-Perez, R. & Enriquez, J. A. The function of the respiratory supercomplexes: their help in IR experiments, and Drs. Iain C. Bruce, Ruiping Xiao, Yingxian Li, and Yan the plasticity model. Biochim. Biophys. Acta 1837, 444–450 (2014). Zhang for valuable comments. This work was supported by the National Key Basic 25. Greggio, C. et al. Enhanced respiratory chain supercomplex formation in Research Program of China (2017YFA0504000, 2016YFA0500403 and 2013CB531200) response to exercise in human skeletal muscle. Cell Metab. 25, 301–311 (2017). and the National Science Foundation of China (31670039, 8182780030, 31821091, 26. Rosca, M. G. et al. Cardiac mitochondria in heart failure: decrease in respirasomes and 31470811). and oxidative phosphorylation. Cardiovasc. Res. 80,30–39 (2008). 27. Runswick, M. J., Fearnley, I. M., Skehel, J. M. & Walker, J. E. Presence of an acyl carrier protein in NADH:ubiquinone oxidoreductase from bovine heart mito- AUTHOR CONTRIBUTIONS chondria. FEBS Lett. 286, 121–124 (1991). – H.C. and X.W. conceived and supervised the study. T.H., R.Z., C.J., and Y.W. performed 28. Hiltunen, J. K. et al. Mitochondrial fatty acid synthesis an adopted set of enzymes experiments and analyzed data with helps from Q.M. and X.H. and S.L. W.D. making a pathway of major importance for the cellular metabolism. Prog. Lipid contributed to the RNA-seq data analysis. X.W. and H.C.wrote the paper with inputs Res. 49,27–45 (2010). from all authors. All the authors participated in data interpretation. 29. Feng, D., Witkowski, A. & Smith, S. Down-regulation of mitochondrial acyl carrier protein in mammalian cells compromises protein lipoylation and respiratory complex I and results in cell death. J. Biol. Chem. 284, 11436–11445 (2009). ADDITIONAL INFORMATION 30. Vinothkumar, K. R., Zhu, J. & Hirst, J. Architecture of mammalian respiratory complex I. Nature 515,80–84 (2014). Supplementary information accompanies this paper at https://doi.org/10.1038/ 31. Floyd, B. J. et al. Mitochondrial protein interaction mapping identifies regulators s41422-019-0208-x. of respiratory chain function. Mol. cell 63, 621–632 (2016). 32. Van Vranken, J. G., et al. The mitochondrial acyl carrier protein (ACP) coordinates Competing interests: The authors declare no competing interests. mitochondrial fatty acid synthesis with iron sulfur cluster biogenesis. Elife 5, https://doi.org/10.7554/eLife.17828 (2016). 33. Stroud, D. A. et al. Accessory subunits are integral for assembly and function of REFERENCES human mitochondrial complex I. Nature 538, 123–126 (2016). fi 1. Neubauer, S. The failing heart–an engine out of fuel. N. Engl. J. Med. 356, 34. Zheng, M. et al. Cardiac-speci c ablation of Cypher leads to a severe form of dilated – 1140–1151 (2007). cardiomyopathy with premature death. Hum. Mol. Genet. 18,701 713 (2009). 2. Ingwall, J. S. ATP and the Heart. (Kluwer Academic Publishers, Norwell, MA, 35. Balaban, R. S., Kantor, H. L., Katz, L. A. & Briggs, R. W. Relation between work and 2002). phosphate metabolite in the in vivo paced mammalian heart. Science 232, – 3. Droge, W. Free radicals in the physiological control of cell function. Physiol. Rev. 1121 1123 (1986). 82,47–95 (2002). 36. Neely, J. R., Rovetto, M. J., Whitmer, J. T. & Morgan, H. E. Effects of ischemia on 4. Beckman, K. B. & Ames, B. N. The free radical theory of aging matures. Physiol. Rev. function and metabolism of the isolated working rat heart. Am. J. Physiol. 225, – 78, 547–581 (1998). 651 658 (1973). 5. DiMauro, S. & Schon, E. A. Mitochondrial respiratory-chain diseases. N. Engl. J. 37. Matthews, P. M., Bland, J. L., Gadian, D. G. & Radda, G. K. The steady-state rate of Med. 348, 2656–2668 (2003). ATP synthesis in the perfused rat heart measured by 31P NMR saturation transfer. – 6. Ingwall, J. S. Transgenesis and cardiac energetics: new insights into cardiac Biochem. Biophys. Res. Commun. 103, 1052 1059 (1981). metabolism. J. Mol. Cell Cardiol. 37, 613–623 (2004). 38. Allue, I. et al. Evidence for rapid consumption of millimolar concentrations of 7. Maack, C. & O’Rourke, B. Excitation-contraction coupling and mitochondrial cytoplasmic ATP during rigor-contracture of metabolically compromised single – energetics. Basic Res. Cardiol. 102, 369–392 (2007). cardiomyocytes. Biochem. J. 319(Pt 2), 463 469 (1996). fl 8. Nicholls, D. G. & Ferguson, S. L. Bioenergetics. (Academic Press, London, 2002). 39. Wang, X. et al. Mitochondrial ashes regulate ATP homeostasis in the heart. Elife 9. Wu, M. et al. Structure of mammalian respiratory supercomplex I1III2IV1. Cell 167, 6, https://doi.org/10.7554/eLife.23908 (2017). 1598–1609 e1510 (2016). 40. Hiltunen, J. K. et al. Mitochondrial fatty acid synthesis and respiration. Biochim. Biophys. Acta 1797, 1195–1202 (2010).

Cell Research (2019) 29:754 – 766 Article 766 41.Jian,C.etal.Deficiency of PHB complex impairs respiratory supercomplex 56. Guerrero-Castillo, S. et al. The assembly pathway of mitochondrial respiratory formation and activates mitochondrial flashes. J. Cell Sci. 130,2620–2630 chain complex I. Cell Metab. 25, 128–139 (2017). (2017). 57. Acin-Perez, R. et al. Respiratory active mitochondrial supercomplexes. Mol. cell 32, 42. Mitsopoulos, P. et al. Stomatin-like protein 2 is required for in vivo mitochondrial 529–539 (2008). respiratory chain supercomplex formation and optimal cell function. Mol. Cell. 58. Medeiros, D. M. Assessing mitochondria biogenesis. Methods 46, 288–294 (2008). Biol. 35, 1838–1847 (2015). 59. Cheng, H., Lederer, W. J. & Cannell, M. B. Calcium sparks: elementary events 43. Wagener, N., Ackermann, M., Funes, S. & Neupert, W. A pathway of protein underlying excitation-contraction coupling in heart muscle. Science 262, 740–744 translocation in mitochondria mediated by the AAA-ATPase Bcs1. Mol. cell 44, (1993). 191–202 (2011). 60. Shang, W. et al. Cyclophilin D regulates mitochondrial flashes and metabolism in 44. Maio, N. & Rouault, T. A. Iron-sulfur cluster biogenesis in mammalian cells: new cardiac myocytes. J. Mol. Cell Cardiol. 91,63–71 (2016). insights into the molecular mechanisms of cluster delivery. Biochim. Biophys. Acta 61. Shen, T. et al. Mitofusin-2 is a major determinant of oxidative stress-mediated 1853, 1493–1512 (2015). heart muscle cell apoptosis. J. Biol. Chem. 282, 23354–23361 (2007). 45. Roche, B. et al. Iron/sulfur proteins biogenesis in prokaryotes: formation, reg- 62. Zhang, S. J. et al. Isoform evolution in primates through independent combina- ulation and diversity. Biochim. Biophys. Acta 1827, 455–469 (2013). tion of alternative RNA processing events. Mol. Biol. Evol. 34, 2453–2468 (2017). 46. Karamanlidis, G. et al. Mitochondrial complex I deficiency increases protein 63. Huang da, W., Sherman, B. T. & Lempicki, R. A. Systematic and integrative analysis acetylation and accelerates heart failure. Cell Metab. 18, 239–250 (2013). of large gene lists using DAVID bioinformatics resources. Nat. Protoc. 4,44–57 47. Murphy, E. & Steenbergen, C. Mechanisms underlying acute protection from (2009). cardiac ischemia-reperfusion injury. Physiol. Rev. 88, 581–609 (2008). 64. Huang da, W., Sherman, B. T. & Lempicki, R. A. Bioinformatics enrichment tools: 48. Yellon, D. M. & Hausenloy, D. J. Myocardial reperfusion injury. N. Engl. J. Med. 357, paths toward the comprehensive functional analysis of large gene lists. Nucleic 1121–1135 (2007). Acids Res. 37,1–13 (2009). 49. Eltzschig, H. K. & Eckle, T. Ischemia and reperfusion–from mechanism to trans- lation. Nat. Med. 17, 1391–1401 (2011). 50. Muller, F. L., Liu, Y. & Van Remmen, H. Complex III releases superoxide to both Open Access This article is licensed under a Creative Commons sides of the inner mitochondrial membrane. J. Biol. Chem. 279, 49064–49073 Attribution 4.0 International License, which permits use, sharing, (2004). adaptation, distribution and reproduction in any medium or format, as long as you give 51. Matsumoto-Ida, M. et al. Real-time 2-photon imaging of mitochondrial function in appropriate credit to the original author(s) and the source, provide a link to the Creative perfused rat hearts subjected to ischemia/reperfusion. Circulation 114, Commons license, and indicate if changes were made. The images or other third party 1497–1503 (2006). material in this article are included in the article’s Creative Commons license, unless 52. Wang, X. et al. Cardioprotection of ischemia/reperfusion injury by cholesterol- indicated otherwise in a credit line to the material. If material is not included in the dependent MG53-mediated membrane repair. Circ. Res. 107,76–83 (2010). article’s Creative Commons license and your intended use is not permitted by statutory 53. Moreno-Lastres, D. et al. Mitochondrial complex I plays an essential role in human regulation or exceeds the permitted use, you will need to obtain permission directly respirasome assembly. Cell Metab. 15, 324–335 (2012). from the copyright holder. To view a copy of this license, visit http://creativecommons. 54. Chouchani, E. T. et al. Ischaemic accumulation of succinate controls reperfusion org/licenses/by/4.0/. injury through mitochondrial ROS. Nature 515, 431–435 (2014). 55. Fiedorczuk, K. et al. Atomic structure of the entire mammalian mitochondrial complex I. Nature 538, 406–410 (2016). © The Author(s) 2019

Cell Research (2019) 29:754 – 766